babillty of error is diminished in the same degree[1]. Let us now substitute for the plate of glass a flake of alum, sugar, or ice; we shall find that the needle of the galvanometer is perfectly at rest: if there is any heat transmitted, it is therefore not more than of the whole radiation. Thus it is true that the transmission of these three substances reduced to plates of 2mm·6 in thickness and exposed to the radiation of a body heated to 390° is null or less than 1200dth part of the whole incident heat. It is by operations analogous to this that I have been able to ascertain the limits of the values of the zeros of transmission.
Now that we know the degree of exactness to which the measures contained in our table have been carried, we may proceed to state the consequences to which they lead.
Let us, for the moment, not notice the results obtained with the rock salt. The order of the transmissions has no relation to the degree of transparency, as we have already determined in our first series of experiments. It is not strictly the same when we change the calorific source; but each substance exposed to the successive action of the four radiations presents a like order of decrease in respect to the quantities which it transmits from each of the sources; that is to say, that all the substances transmit quantities of heat which are feeble in proportion as the temperature of the radiating source is low. There are several cases in which the transmissions are nothing; but these cases do not make
- ↑ This mode of estimating the energy of the calorific radiations enables us to determine without difficulty the ratios existing between the arcs described by the magnetic needle of the galvanometer and the corresponding forces. Let us suppose the calorific source removed sufficiently far from the pile to produce but a feeble deviation of the galvanometer; one of 10°, for example. In the passage of the calorific rays let there be interposed a plate which transmits a certain fraction of the incident heat. We shall suppose this fraction to be 12; the needle will descend to 2°. By bringing the source near, the deviation produced through the plate will be increased. Let us stop, when the needle shall have reached 4°, 6°, 8°, &c. successively; the calorific source will then emit upon the pile twice, thrice, or four times as much heat as before; for the transmission through the same plate exposed to a constant source of heat is always in a constant ratio, and the forces of deviation are proportional to the degrees in those arcs that are very near zero. Let the force which causes the galvanometer to describe the first degree of the scale be represented as 1, we shall then have 10 for the first force or quantity of incident heat, 20 for the second, 30 for the third, 40 for the fourth, &c. Now we know that the first force answers to 10°. In order to determine the deviation produced by the force 20 we have only to remove the plate when the galvanometer points to 4°; the calorific rays will then fall immediately on the pile, the angle of deviation will increase, and if the proportionality of the degrees to the forces continues through the whole extent of the arc of the first 20 degrees we shall see the index stop at 20°: at all events we shall have the corresponding indication. By repeating the same operation when the galvanometer points to 6°, 8°, we shall obtain the quantities sought, that is to say, the degrees answering to the forces 20, 30, 40, &c. Thus we may verify the results contained in the tables of intensities already made, or determine the elements necessary for the construction of new tables.
